set Resource_List.nodes to "1:ppn=8" (without the quotes) and it should work. At least it does for me on queen bee.
As far as we know there is no way to automatically figure out how the nodes string has to be. (Let us know if you know anything)
But of course you should have gotten the error: ** Must specify :ppn=8 on nodes resource **
I will fix this next.
Roland
On Tue, Jul 20, 2010 at 4:21 PM, Steven R. Brandt <sbrandt@xxxxxxxxxxx> wrote:
This is on queen bee (qb.loni.org, a Teragrid machine), and yes qsub is
in the path (/usr/local/bin/qsub). It is a wrapper script around the
binary.
Cheers,
Steve
On 07/20/2010 01:52 PM, Roland Schulz wrote:
Hi,
yes sorry the PBS support in 4.0 is lacking good error
reporting. Sorry.
Is the PATH variable set to where qsub is found? What are the
required qsub flags and/or could you provide a sample script? Is the
qsub the default torque qsub binary or is it a wrapper script around
the qsub binary? What cluster do you try to submit on (if it is part of
e.g. teragrid I might be able to test it myself)?
Roland
On Tue, Jul 20, 2010 at 10:18 AM, Steven R.
Brandt <sbrandt@xxxxxxxxxxx>
wrote:
I
attempted to set up a remote PBS resource using the new Helios build.
Configuration went smoothly enough. I set up port forwarding, clicked
start resource, and
saw lots of green.
Unfortunately, I can't submit jobs. It fails with an empty message and
no stack trace (see attached
image).
In my log I see this:
!ENTRY org.eclipse.ptp.launch 4 0 2010-07-20 09:16:10.201
!MESSAGE
